© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 1581
    Prev
    Next
    INDEX
    Explanations

    suggestions for community activities or gatherings

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    672
    +0.13
    0.6%
    335
    +0.10
    0.5%
    508
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.13
    0.06
    672
    +0.10
    0.04
    105
    +0.10
    0.04
    Negative Logits
     settlements
    -0.73
     att
    -0.71
     denies
    -0.68
    hovah
    -0.67
    xit
    -0.65
     UNCLASSIFIED
    -0.64
     deposited
    -0.64
     exits
    -0.64
    )</
    -0.64
     desp
    -0.63
    POSITIVE LOGITS
     automate
    0.83
     quantify
    0.81
     recreate
    0.80
     brainstorm
    0.80
     themed
    0.77
     showcase
    0.75
     explore
    0.75
     Kickstarter
    0.75
     feasibility
    0.73
     DIY
    0.71
    Activations Density 0.495%

    No Known Activations